The assertive community treatment (ACT) model has been criticized as being highly coercive. It is critical that these claims be systematically studied. The proposed study will use a cross-sectional design to examine the use of coercion on 25 teams who are implementing the ACT model in Indiana; teams vary in their degree of fidelity to the ACT model. Coercion will be assessed with three newly developed independent measures: objective measures of teams' position of control over clients (e.g., percentage of clients on involuntary commitments, with representative payeeships, or on depot injection), practioners' self-report of coercive practices used with clients, and observer ratings of teams' use of coercion. Several predictors of coercion will also be examined (i.e., ACT fidelity, quality of assessment/treatment planning/treatment, practitioners' level of training and attitudes, and supervision). The proposed study will introduce alternative methods for measuring coercion, provide much-needed base rate information about teams' use of coercion, and examine possible predictors of coercion, which could direct future research and development of targeted interventions to reduce the use of coercion. [unreadable] [unreadable]